15213773, 2022, 35, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/anie.202207002 by Spanish Cochrane National Provision (Ministerio de Sanidad), Wiley Online Library on [18/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Angewandte Research Articles Chemie www.angewandte.org How to cite: Angew. Chem. Int. Ed. 2022, 61, e202207002 Thermoelectrics Very Important Paper International Edition: doi.org/10.1002/anie.202207002 German Edition: doi.org/10.1002/ange.202207002 Surface Functionalization of Surfactant-Free Particles: A Strategy to Tailor the Properties of Nanocomposites for Enhanced Thermoelectric Performance Cheng Chang, Yu Liu, Seung Ho Lee, Maria Chiara Spadaro, Kristopher M. Koskela, Tobias Kleinhanns, Tommaso Costanzo, Jordi Arbiol, Richard L. Brutchey, and Maria Ibáñez* Introduction Abstract: The broad implementation of thermoelectric- ity requires high-performance and low-cost materials. Thermoelectric materials directly and reversibly transform One possibility is employing surfactant-free solution heat and electricity. Considering heat as the most common synthesis to produce nanopowders. We propose the form of waste energy, this phenomenon has been explored strategy of functionalizing “naked” particles’ surface by as a potential source of clean, non-fossil energy.[1] The inorganic molecules to control the nanostructure and, conversion efficiency of thermoelectric materials is deter- consequently, thermoelectric performance. In particular, mined by the dimensionless figure of merit, z T = S2 σ T/k, we use bismuth thiolates to functionalize surfactant-free where S, σ, k, and T are the Seebeck coefficient, electrical SnTe particles’ surfaces. Upon thermal processing, conductivity, thermal conductivity, and absolute temper- bismuth thiolates decomposition renders SnTe-Bi2S3 ature, respectively. During the last century, material scien- nanocomposites with synergistic functions: 1) carrier tists have tried to maximize materials’ z T to enhance concentration optimization by Bi doping; 2) Seebeck conversion efficiency. However, the complex and predom- coefficient enhancement and bipolar effect suppression inantly counterproductive relationships between the param- by energy filtering; and 3) lattice thermal conductivity eters determining z T make it challenging.[2] reduction by small grain domains, grain boundaries and The most successful strategies to enhance performance nanostructuration. Overall, the SnTe-Bi2S3 nanocompo- have been based on refining the materials’ band structures sites exhibit peak z T up to 1.3 at 873 K and an average and scattering centers simultaneously, which require superb z T of � 0.6 at 300–873 K, which is among the highest control over the material structural properties.[3] Defects reported for solution-processed SnTe. ranging from atomic impurities to embedded nanoparticles must be precisely engineered to maximize electron transport while hampering phonon movement. Inorganic nanocompo- sites are among the most suitable materials to tackle such endeavors. There are plenty of methods to produce inorganic nanocomposites, the most common in thermo- electric materials being spinodal decomposition in high- [*] Dr. C. Chang, Dr. Y. Liu, S. Ho Lee, M. Chiara Spadaro, temperature reactions. This synthetic strategy delivered T. Kleinhanns, Dr. T. Costanzo, Prof. M. Ibáñez record z Ts in different material systems, e.g. PbTe, SnTe, or Institute of Science and Technology Austria SiGe.[4] While successful, it requires large amounts of Am Campus 1, 3400 Klosterneuburg (Austria) energy, long reaction time, and ultra-high purity reagents, E-mail: mibanez@ist.ac.at therefore, yielding materials that are too expensive for large Prof. J. Arbiol scale implementation. Catalan Institute of Nanoscience and Nanotechnology (ICN2) Alternatively, in the last decade, solution processing CSIC, and BIST, 08193 Barcelona, Catalonia (Spain) methods have been explored to reduce material production Prof. J. Arbiol costs. Solution syntheses commonly use low-cost reagents, ICREA low temperatures, and short reaction time to produce Pg. Lluís Companys 23, 08010 Barcelona, Catalonia (Spain) powders. The most widely used method to produce powders K. M. Koskela, Prof. R. L. Brutchey in solution for thermoelectric materials is the so-called Department of Chemistry, University of Southern California surfactant-free since it has high reaction yields and renders Los Angeles, CA 90089 (USA) particles without insulating long-chain organic molecules, © 2022 The Authors. Angewandte Chemie International Edition typically found when surfactants are used.[5] published by Wiley-VCH GmbH. This is an open access article under the terms of the Creative Commons Attribution License, which From particles to dense materials, sintering is permits use, distribution and reproduction in any medium, provided necessary.[6] The sintering process involves densification and the original work is properly cited. particle coarsening, both of which will determine the micro- Angew. Chem. Int. Ed. 2022, 61, e202207002 (1 of 7) © 2022 The Authors. Angewandte Chemie International Edition published by Wiley-VCH GmbH 15213773, 2022, 35, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/anie.202207002 by Spanish Cochrane National Provision (Ministerio de Sanidad), Wiley Online Library on [18/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Angewandte Research Articles Chemie structure of the nanocomposite.[7] Particle surfaces play a To form SnTe-Bi2S3 nanocomposites, SnTe NPs were crucial role in all these phenomena and therefore modifying dispersed in N-methyl formamide, then different amounts their characteristics is a powerful strategy to engineer (x mol %) of dissolved Bi2O3 or Bi2S3 in a thiol-amine nanocomposite structural properties. solution (1,2-ethanedithiol, EDT; ethylenediamine, en) were One possibility, barely explored, is the particle surface added. The mixtures were vigorously stirred at room functionalization with different organic or inorganic mole- temperature in an N2 filled glovebox for 24 h. The treated cules. However, the high temperatures employed, usually SnTe NPs were then precipitated from the solution, washed above 400 °C, will decompose those surface species. While with acetone thoroughly, and dried under vacuum for this can be seen as a hurdle, it also represents an further processing into pellets following the same steps as opportunity. The decomposition products can be used to for the bare SnTe NPs, resulting in dense SnTe-Bi2S3 control the matrix composition or even to form different nanocomposites (Figure 1). The nanocomposites produced types of secondary phases. Herein, we demonstrate that with Bi2O3 solutions are referred to as SnTe-x %Bi2S3, where adding surface molecules to “naked” particles and tailoring x = 0.5, 1, 1.5, 2, 2.5, and 3. The nanocomposites produced their decomposition is a powerful tool for precisely control- with Bi2S3 solutions are referred to as SnTe-x %*Bi2S3*, ling nanocomposite structural and thermoelectric properties where x = 1, 2, and 3. in SnTe. In particular, we prepare bismuth thiolates in a The X-ray diffraction (XRD) patterns of the consoli- thiol-amine mixture. These molecular solutes are then used dated SnTe-Bi2S3 nanocomposites produced with the par- to functionalize surfactant-free SnTe particles’ surfaces. The ticles treated with Bi2O3 or Bi2S3 solutions match the single surface-modified SnTe particles are then subjected to phase of the SnTe rock-salt structure without any visible various thermal processes that yield dense SnTe-Bi2S3 nano- crystalline impurity phases detected (Figure S2). In the case composites. Depending on the precursors used to prepare of treating SnTe NPs with the Bi2O3 solution, the lattice the bismuth thiolates, different decomposition products are parameter of SnTe increased with the Bi2O3 molar content obtained, resulting in nanocomposites with distinct structur- used to treat the SnTe NPs (Figure 2a). On the contrary, the al properties. The produced SnTe-Bi2S3 nanocomposites SnTe lattice parameters of the nanocomposites prepared exhibit peak z T up to 1.3 at 873 K and an average z T of with Bi2S3 solution remain almost unchanged despite � 0.6 at 300–873 K, which is among the highest reported for increasing the Bi2S3 content. To understand the differences SnTe. between the composites prepared using Bi2O3 and Bi2S3, we looked at the identity of the molecular solutes for solutions of Bi2O3 and Bi2S3 dissolved in 1 : 10 (vol/vol) mixtures of Results and Discussion EDT and en at room temperature and ambient pressure. In particular, we performed direct-injection electrospray ion- Ligand-free SnTe nanoparticles (NPs) were prepared in ization mass spectrometry (ESI-MS) using a liquid chroma- water at ambient pressure using SnCl2 and Te as precursors. tography quadrupole time-of-flight mass spectrometer (LC/ This synthetic method was chosen due to its simplicity and Q-TOF-MS) and UV/Vis spectroscopy measurements. Fig- uses very inexpensive precursors. The resulting SnTe NPs ure S3 shows the negative ion mode mass spectra for Bi2O3 have an irregular shape with dimensions of 20–50 nm (Fig- and Bi2S3 solutions, respectively. The three most abundant ure S1). The as-synthesized SnTe NPs were purified from peaks at m/z = 272.9, 332.9, and 392.9 match well with the the reaction byproducts by precipitation, rinsing with calculated values for [BiS2] , [BiS(C2H4S2)] (or BiS- ethanol and acetone. The washed NPs were then dried (EDT) ), and [Bi(C2H4S2)2] (or Bi(EDT)2 ), respectively, under vacuum overnight at room temperature, annealed in for both samples with the other peaks matching to those of forming gas (95 % N2 + 5 % H2), and consolidated into the solvent blank (Table 1). Positive-ion mode mass spectra pellets under vacuum using the spark plasma sintering peaks most likely correspond to various alkylammonium technique. Further details of the entire process can be found species as reported previously for other thiol-amine in the Supporting Information. solutions.[8] UV/Vis absorption spectroscopy was used to Figure 1. SnTe-Bi2S3 nanocomposite synthetic process: SnTe NPs are capped with bismuth thiolates, then washed, dried, annealed, and consolidated into a pellet by applying temperature (T) and pressure (P), forming the SnTe-Bi2S3 nanocomposites. Angew. Chem. Int. Ed. 2022, 61, e202207002 (2 of 7) © 2022 The Authors. Angewandte Chemie International Edition published by Wiley-VCH GmbH 15213773, 2022, 35, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/anie.202207002 by Spanish Cochrane National Provision (Ministerio de Sanidad), Wiley Online Library on [18/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Angewandte Research Articles Chemie Figure 2. a) The SnTe lattice parameter of SnTe-Bi2S3 nanocomposite produced with Bi2O3/Bi2S3 solutions. b) Pictures of the Bi2O3/Bi2S3 solutions and the resulting products after annealing at 300 °C under vacuum. c) The XRD patterns of recrystallization products of Bi2O3/Bi2S3 solutions. Table 1: Molecular formulas and proposed structures for ions observed in negative-ion mode mass spectra. Measured Molecular Proposed structures Calculated m/z formula m/z 272.9 [BiS2] 273.1 332.9 [BiS(C2H4S2)] 333.2 392.9 [Bi(C2H4S2)2] 393.3 further investigate the absorption signatures of these molec- variation in SnTe lattice parameter to the differences seen in ular solutes. Dilute solutions of both Bi2O3 and Bi2S3 exhibit the decomposition products of the two solutions. During the an absorption band at λmax = 392 nm (Figure S4) that is thermal processing in the nanocomposite produced with the assigned to a ligand-to-metal charge transfer band (LMCT), Bi2O3 solution, we hypothesize that Bi produced during the consistent with other experimentally observed and computa- decomposition diffuses into SnTe lattice occupying Sn or Sn tionally predicted bismuth(III) thiolate complexes.[9] This vacancy sites, leading to the lattice expansion. Thus, BiIII suggests that both Bi2O3 and Bi2S3 yield similar bismuth should also act as an aliovalent dopant modifying the charge thiolate solutes upon dissolution, corroborating the conclu- carrier concentration, as is confirmed later on. sions drawn from ESI-MS measurements. While indirect characterization methods such as those used here (ESI-MS, UV/Vis) point to mostly identical Nanocomposite Structural Properties solutes resulting from the two precursors, we found distinct differences after recrystallizing them by annealing under Transmission electron microscopy (TEM) related analysis of vacuum at 300 °C. The XRD patterns reveal that Bi2O3 the surface-treated NPs showed a ca. 10 nm amorphous shell solution transforms into a biphasic Bi2S3 and metallic Bi around the SnTe NPs constituted of homogeneously dis- mixture, whereas the Bi2S3 solution yields phase-pure Bi2S3 tributed Bi and S atoms (Figures 3a, b). We attribute this (Figures 2b, c).[10] We hypothesize that the differences during amorphous shell to the adsorption of the different bismuth crystallization of both solutions comes from the evolution of thiolates: [BiS2] , [BiS(C2H4S2)] and [Bi(C2H4S2)2] . stoichiometric amounts of H2O from the liberation of lattice To evaluate the microstructure of the nanocomposites, oxygen in Bi2O3, as seen in previous studies.[11] the SnTe-1.5 %Bi2S3 pellet was thinned to electron trans- parency by Ar + polishing to produce a self-suspended e:g: Bi2 O3 þ 4 C2 H6 S2 ! 2 ½BiðC2 H4 S2 Þ2 � þ 2 Hþ lamella. Figure 3c shows the linear energy-dispersive X-ray (1) spectrometry (EDS) scan along the selected grain boundary þ3 H2 O (white arrow). Quantitative results reveal evident Sn/Te Since both Bi2O3 and Bi2S3 solutions are constituted of spectrum valleys and Bi/S spectrum peaks at the grain the same molecular species (Table 1), we attributed the boundary areas, indicating that during the thermal process- Angew. Chem. Int. Ed. 2022, 61, e202207002 (3 of 7) © 2022 The Authors. Angewandte Chemie International Edition published by Wiley-VCH GmbH 15213773, 2022, 35, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/anie.202207002 by Spanish Cochrane National Provision (Ministerio de Sanidad), Wiley Online Library on [18/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Angewandte Research Articles Chemie Figure 3. a) STEM image of a SnTe NP coated with bismuth thiolate complexes from the dissolution of Bi2O3 in the alkahest solvent, and b) its corresponding EDS maps. c) STEM of SnTe-1.5 %Bi2S3 nanocomposite produced with the Bi2O3 solution. The inset image is the linear EDS spectrum along the white arrow direction. d) The corresponding EDS maps. e) HRTEM image of SnTe-1.5 %Bi2S3 nanocomposite produced with Bi2O3 solution. Inset, the phase-filtered structural map shows the SnTe phase marked as green and blue with different orientations and Bi2S3 marked as red. The inset top right shows the crystallographic mapping of two adjacent SnTe grains with a Bi2S3 nanoprecipitate in between. ing, the bismuth complexes crystallize, forming Bi2S3 secon- dary phases because of the low solubility of Bi into SnTe.[12] Compositional maps show that Bi2S3 is mostly located along the grain boundaries forming grain boundary complexions and in the form of Bi2S3 nanodots (Figure 3d). High- resolution transmission electron microscopy (HRTEM) verified the presence of Bi2S3 nanoprecipitates with an orthorhombic crystal structure between SnTe crystalline domains (Figure 3e). In particular, the marked red region corresponds to the Bi2S3 visualized along the [001] zone axis and the selected green region corresponds to the SnTe visualized along the [100] zone axis parallel to the electron beam direction. In the inset of Figure 3e, the corresponding phase-filtered structural mapping is shown, where the SnTe phase from two adjacent grains is marked as green and blue, and the Bi2S3 phase is displayed in red. The Bi2S3 phase lies between the two misoriented adjacent SnTe grains, consis- tent with the EDS results. Additional STEM, SEM, EDS, and HRTEM data can be found in Figures S5–S15 and Table S1. Figure 4. SEM image of a) bare SnTe, b) SnTe-1.5 %Bi2S3 nanocompo- Figure 4 compares the scanning electron microscopy site produced with the Bi2O3 solution. c) Grain size distribution and corresponding Gaussian distribution fitting, and d) schematic of grain images of the pellets produced with bare SnTe NPs and growth inhibition in the SnTe-Bi2S3 nanocomposite. those treated with 1.5 % Bi2O3 solution. SEM images of the other SnTe-x %Bi2S3 nanocomposites formed with different amounts of Bi2O3 solution can be found in Figure S8. SnTe- 1.5 %Bi2S3 nanocomposite contains smaller grains than bare variance also reduces from 1.23 to 0.7 μm (Figure 4c). We SnTe. The average grain size decreases from 1.17 μm in bare associate these with the presence of Bi2S3 secondary phase SnTe to 0.74 μm in SnTe-1.5 %Bi2S3 nanocomposite, and the as nanoprecipitates at the grain boundaries. Angew. Chem. Int. Ed. 2022, 61, e202207002 (4 of 7) © 2022 The Authors. Angewandte Chemie International Edition published by Wiley-VCH GmbH 15213773, 2022, 35, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/anie.202207002 by Spanish Cochrane National Provision (Ministerio de Sanidad), Wiley Online Library on [18/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Angewandte Research Articles Chemie Grains grow through atomic diffusion across the grain Bare SnTe and all the produced SnTe-x %Bi2S3 nano- boundaries, which can be envisioned as grain boundary composites have a positive S, which increases with temper- movement.[13] The presence of Bi2S3 at the grain boundary ature (Figure 5b). S values of all SnTe-Bi2S3 nanocomposites inhibits grain growth due to the stagnant diffusion of SnTe approach the same maximum value of � 180 μV K 1 at through Bi2S3. Moreover, Bi2S3 nanoprecipitates act as 873 K. The room-temperature S increases with x from pinning centers hindering grain boundary movement, a � 35 μV K 1 for bare SnTe to � 110 μV K 1 for SnTe-3 % phenomenon known as Zenner pinning (Figure 4d).[14] As a Bi2S3 nanocomposite. The combination of the optimized σ result, mass diffusion is kinetically stabilized, hindering grain and enhanced S leads to high power factors (PF = σS2) in growth. The same effect can be observed in SnTe-Bi2S3 SnTe-Bi2S3 nanocomposites (Figure 5c). Surprisingly, the nanocomposites formed with the Bi2S3 solution (Figure S9). composites with 2.5 % and 3 %Bi2S3 exhibit a steady high PF within the whole temperature range studied, mainly deriving from the high S at low temperatures. The same phenomenon Thermoelectric Properties happens for SnTe-3 %*Bi2S3* as well. To get an insight into the high S, we investigated the All samples exhibit heavily doped semiconductor behavior, room-temperature S of SnTe-Bi2S3 nanocomposites as a with σ decreasing with rising temperatures (Figure 5a). SnTe function of pH, and compared it with bare, Bi-, Sb-, Cd-, and is intrinsically p-type due to its off-stoichiometric nature In-doped SnTe from previous works (Figure 5d).[16] The with a large number of Sn lattice vacancies that yields very theoretical Pisarenko relationship was calculated based on a high intrinsic doping levels.[15] Noticeably, the room temper- two-valence-band model.[17] S values of Cd- and In-doped ature σ decreases from � 6100 S cm 1 for bare SnTe to SnTe lie above the calculated Pisarenko line due to the band � 3500 S cm 1 for SnTe-3 %Bi2S3 nanocomposites with the convergence effect and resonant doping, respectively.[18] On carrier concentration (pH) reduces from 1.4 × 1021 cm 3 to the other hand, S values of the previously reported Bi-and 3.0 × 1019 cm 3. Such electrical conductivity and carrier con- Sb-doped SnTe fit well with the estimated Pisarenko, centration changes are associated with Bi-doping, which acts indicating Bi and Sb make no noticeable difference in the as an electron donor, reducing the overall hole concentra- SnTe band structure.[19] However, in this work, all S values tion. The nanocomposite prepared with 1 mol % Bi2S3 of SnTe-Bi2S3 nanocomposites lie far above the Pisarenko solution referred to as SnTe-1 %*Bi2S3* is also presented for line. We attribute this to the energy filtering occurring due comparison. As expected, σ in SnTe-1 %*Bi2S3* is much to the presence of Bi2S3 rich grain boundaries found in the higher than SnTe-1 % Bi2S3 due to the absence of Bi doping TEM study. Figure S17 illustrates the relative band structure evidenced by the lattice parameters. The data corresponding energy of SnTe and Bi2S3 to vacuum.[20] The discrepant band to all the SnTe-x %*Bi2S3* nanocomposites can be found in structures between SnTe and Bi2S3 establish effective Figures S16. potential barriers in which low-energy charge carriers are filtered out. In contrast, high-energy charge carriers can pass through the barrier freely, leading to higher DOS effective mass (m*). We calculated m* of all SnTe-Bi2S3 nano- composites, according to m* � me(μw/μH)2/3 (Hall mobility μH, weighted mobility μw and m* can be found in Table S2 and Figure S18),[21] confirming that m* in SnTe-Bi2S3 nano- composites is higher than that of bare SnTe. Apart from the energy filtering effect, the potential barrier is also respon- sible for the carrier concentration reduction since low energy charge carriers are trapped in the potential well, which is confirmed in SnTe-*Bi2S3* nanocomposites where carrier concentration still decrease without Bi doping (Table S2).[22] Aside from the enlarged PF, the surface functionaliza- tion with bismuth thiolates also provides effective means to reduce k. Figure 6a shows the total thermal conductivity (ktot) and lattice thermal conductivity (klat) as a function of temperature (specific heat Cp, thermal diffusivity α, Lorentz number L, sample density ρ and electronic thermal con- ductivity kele can be found in Figure S19, and Table S3. L is calculated using the single parabolic band (SPB) model with acoustic phonon scattering, which is also discussed in Figure 5. The electrical transport properties of SnTe-x %Bi2S3 nano- composites produced with the Bi2O3 solution. a) Electrical conductivity. Supporting Information). The room-temperature ktot of bare b) Seebeck coefficient. c) Power factor. d) The Pisarenko plot of SnTe SnTe is � 7 W m 1 K 1, arising from the undesirable high kele. according to the two-band Kane model in reference [17]. The Seebeck Using Bi2O3 solutions can significantly suppress the high kele coefficient of undoped/Bi-/Sb-/In-/Cd- doped SnTe using the traditional by reducing pH (Table S2) and introduce additional scatter- melting method are taken from reference [16]. ing sources. Angew. Chem. Int. Ed. 2022, 61, e202207002 (5 of 7) © 2022 The Authors. Angewandte Chemie International Edition published by Wiley-VCH GmbH 15213773, 2022, 35, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/anie.202207002 by Spanish Cochrane National Provision (Ministerio de Sanidad), Wiley Online Library on [18/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Angewandte Research Articles Chemie effect onset.[25a] However, k values of all SnTe-Bi2S3 nano- composites decrease monotonically in the whole temper- ature range (Figure 6c), suggesting the suppressed bipolar thermal conductivity. To further illustrate the potential bipolar effect on k, we investigated ktot kele as a function of temperature. Without the bipolar effect, klat should equal to ktot kele, values and klat and T 1 should satisfy a linear relationship according to the acoustic phonon scattering assumption. Otherwise, ktot kele curve will exhibit an upturn at the low 1000/T region (high temperatures), same as bare SnTe synthesized by melting method shown with a continu- ous black line.[25a] Notably, ktot kele of bare SnTe and SnTe- Bi2S3 nanocomposites in this work continue to decrease upon 673 K (1.48 K 1 in Figure 6c), indicating the suppressed bipolar effect. We attributed this to the potential barriers at the SnTe-Bi2S3 interface, which hinder the thermal-activated electron transport in the conduction band at high temperatures.[26] Collectively, the combination of high PF and low k leads Figure 6. a) The total thermal conductivity (solid symbol) and lattice thermal conductivity (hollow symbol) of SnTe-x %Bi2S3 nanocomposites to a high maximum z T value of � 1.3 at 873 K, Figure 6d, produced with the Bi2O3 solution and SnTe-1 %*Bi2S3* prepared with outperforming any other SnTe nanocomposites synthesized the Bi2S3 solution. b) The lattice thermal conductivity as a function of by the bottom-up assembly, yet using a simple, scalable and Bi2S3 amount for the SnTe-x %Bi2S3 nanocomposites. The dashed black cheap aqueous synthetic method. Also, room-temperature lines represent the Klemens model. c) The (ktot kele) as a function of z T is significantly improved from � 0.04 in bare SnTe to 1000/T. d) The thermoelectric figure of merit z T. The inset table reveals � 0.3 in SnTe-3 %Bi2S3 nanocomposite, leading to a remark- the average z T at 300–873 K. able average z T upgrade in 300–873 K, inset of Figure 6d. Furthermore, the synthetic procedure reported here allows for stable performance upon cycling and operating at high We estimated klat according to the Wiedemann–Franz temperatures. The reproducibility measurements of SnTe- relationship (klat = ktot LσT) (Figure 6a). klat decreases from Bi2S3 nanocomposites can be found in Figure S20 and S21. � 2.5 W m 1 K 1 for bare SnTe to � 1 W m 1 K 1 for SnTe- Bi2S3 nanocomposites at room temperature and decreases with rising temperatures, approaching the amorphous limit Conclusion at 973 K.[23] We carried out Klemens model simulations to evaluate In summary, we presented a simple, scalable and cheap the origin of the low klat in SnTe-Bi2S3 nanocomposites aqueous synthetic method to produce SnTe-Bi2S3 nano- (Detailed calculations can be found in Supporting composites by functionalizing the particle surface with Information).[24] Figure 6b shows the calculated room-tem- bismuth thiolates. The SnTe nanocomposite structural perature klat of SnTe-x %Bi2S3 nanocomposites and the properties are determined by the bismuth thiolates’ decom- simulated by the Klemens model as a function of x. As for position products during the thermal processing. SnTe-Bi2S3 the Klemens model, klat reveals an apparent decrease due to nanocomposites contain smaller SnTe grain domains than the point defects introduced by Bi2S3, assuming Bi2S3 is fully bare SnTe, Bi2S3 nanoprecipitates, and Bi2S3 rich grain alloyed with SnTe. However, the experimental klat values in boundary complexions. These features allow for a syner- SnTe-Bi2S3 nanocomposites are much lower than the calcu- gistic optimization of both electrical and thermal transport lated values, indicating other potential phonon scattering properties. As a result, the peak and average z T of SnTe- sources. We attribute the significant discrepancy between Bi2S3 nanocomposites are improved up to 1.3 at 873 K and experimental and predicted values to the combined phonon � 0.6 at 300–873 K, respectively. These findings pave a scattering at the grain boundaries, Bi2S3 rich grain boundary promising and versatile path to tune material structural complexion, and Bi2S3 nanoprecipitates. These accumulated properties taking advantage of the particles’ surface. By defects, ranging from nanometers to micrometers, scatter functionalizing the surface with an adequate molecule, phonons with corresponding mean free paths effectively. unique nanocomposites can be made to improve thermo- Together with the phonon scattering at the atomic scale, the electric performance. surface treatment allows producing nanocomposites with scattering features at all length scales, which is well-known as all-scale hierarchical phonon scattering.[3d] Acknowledgements SnTe is a narrow band gap semiconductor; hence bipolar conduction has a detrimental effect on the thermal This research was supported by the Scientific Service Units transport.[25] The rise of ktot and klat at � 673 K in bare SnTe (SSU) of IST Austria through resources provided by prepared by the melting method are indicative of the bipolar Electron Microscopy Facility (EMF) and the Nanofabrica- Angew. Chem. Int. Ed. 2022, 61, e202207002 (6 of 7) © 2022 The Authors. Angewandte Chemie International Edition published by Wiley-VCH GmbH 15213773, 2022, 35, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/anie.202207002 by Spanish Cochrane National Provision (Ministerio de Sanidad), Wiley Online Library on [18/10/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Angewandte Research Articles Chemie tion Facility (NNF). This work was financially supported by Snyder, B. R. Ortiz, E. S. Toberer, Appl. Phys. Rev. 2018, 5, IST Austria and the Werner Siemens Foundation. C.C. 021303. acknowledges funding from the FWF “Lise Meitner Fellow- [7] S. J. L. Kang, Sintering: Densification, Grain Growth and Microstructure, Vol. 1, Elsevier Butterworth-Heinemann, Ox- ship” grant agreement M 2889-N. Lise Meitner Project ford, 2004. (M2889-N). Y.L. acknowledges funding from the European [8] a) K. M. Koskela, M. J. Strumolo, R. L. Brutchey, Trends Union’s Horizon 2020 research and innovation program Chem. 2021, 3, 1061–1073; b) X. Zhao, S. D. Deshmukh, D. J. under the Marie Sklodowska-Curie grant agreement No. Rokke, G. Zhang, Z. Wu, J. T. Miller, R. Agrawal, Chem. 754411. R.L.B. thanks the National Science Foundation for Mater. 2019, 31, 5674–5682; c) P. Murria, C. K. Miskin, R. support under DMR-1904719. MCS acknowledge MINECO Boyne, L. T. Cain, R. Yerabolu, R. Zhang, E. C. Wegener, J. T. Miller, H. I. Kenttamaa, R. Agrawal, Inorg. Chem. 2017, Juan de la Cierva Incorporation fellowship (JdlCI 2019) and 56, 14396–14407. Severo Ochoa. M.C.S. and J.A. acknowledge funding from [9] a) Y. He, Y. Bin, Y. Liang, J. Xiang, Comput. Theor. Chem. Generalitat de Catalunya 2017 SGR 327. ICN2 is supported 2012, 994, 91–96; b) H. Sommer, A. Eichhöfer, D. Fenske, Z. by the Severo Ochoa program from Spanish MINECO Anorg. Allg. Chem. 2008, 634, 436–440. (Grant no. SEV-2017-0706) and is funded by the CERCA [10] C. L. McCarthy, D. H. Webber, E. C. Schueller, R. L. Brutch- Programme/Generalitat de Catalunya. This study was sup- ey, Angew. Chem. Int. Ed. 2015, 54, 8378–8381; Angew. Chem. ported by MCIN with funding from European Union 2015, 127, 8498–8501. [11] K. M. Koskela, S. J. Quiton, S. M. Sharada, T. J. Williams, NextGenerationEU (PRTR-C17.I1) and Generalitat de R. L. Brutchey, Chem. Sci. 2022, 13, 3208–3215. Catalunya. [12] K. Adouby, A. A. Touré, G. Kra, J. Olivier-Fourcade, J. C. Jumas, C. P. Vicente, C. R. Acad. Sci. Ser. IIc 2000, 3, 51–58. [13] G. Gottstein, L. S. Shvindlerman, Grain Boundary Migration Conflict of Interest in Metals: Thermodynamics, Kinetics, Applications, Vol. 126, CRC, Boca Raton, 2000. The authors declare no conflict of interest. [14] P. A. Manohar, M. Ferry, T. Chandra, ISIJ Int. 1998, 38, 913– 924. [15] a) Y. Zhang, J. Sun, J. Shuai, X. Tang, G. Tan, Mater. Today Phys. 2021, 19, 100405; b) W. Li, L. Zheng, B. Ge, S. Lin, X. Data Availability Statement Zhang, Z. Chen, Y. Chang, Y. Pei, Adv. Mater. 2017, 29, 1605887; c) J. Tang, B. Gao, S. Lin, J. Li, Z. Chen, F. Xiong, The data that support the findings of this study are available W. Li, Y. Chen, Y. Pei, Adv. Funct. Mater. 2018, 28, 1803586. from the corresponding author upon reasonable request. [16] G. Tan, F. Shi, J. W. Doak, H. Sun, L. D. Zhao, P. Wang, C. Uher, C. Wolverton, V. P. Dravid, M. G. Kanatzidis, Energy Environ. Sci. 2015, 8, 267–277. Keywords: Thermoelectrics · SnTe · Grain Boundary · Alkahest [17] Q. Zhang, B. Liao, Y. Lan, K. Lukas, W. Liu, K. Esfarjani, C. Solutions · Surface Functionalization Opeil, D. Broido, G. Chen, Z. Ren, Proc. Natl. Acad. Sci. USA 2013, 110, 13261–13266. [18] G. Tan, L. D. Zhao, F. Shi, J. W. Doak, S. H. Lo, H. Sun, C. [1] G. Tan, L. D. Zhao, M. G. Kanatzidis, Chem. Rev. 2016, 116, Wolverton, V. P. Dravid, C. Uher, M. G. Kanatzidis, J. Am. 12123–12149. Chem. Soc. 2014, 136, 7006–7017. [2] S. Ortega, M. Ibáñez, Y. Liu, Y. Zhang, M. V. Kovalenko, D. [19] L. D. Zhao, X. Zhang, H. Wu, G. Tan, Y. Pei, Y. Xiao, C. Cadavid, A. Cabot, Chem. Soc. Rev. 2017, 46, 3510–3528. Chang, D. Wu, H. Chi, L. Zheng, S. Gong, C. Uher, J. He, [3] a) Y. Pei, X. Shi, A. Lalonde, H. Wang, L. Chen, G. J. Snyder, M. G. Kanatzidis, J. Am. Chem. Soc. 2016, 138, 2366–2373. Nature 2011, 473, 66–69; b) M. Ibáñez, Z. Luo, A. Genc, L. [20] a) J. Nishitani, K. M. Yu, W. Walukiewicz, Appl. Phys. Lett. Piveteau, S. Ortega, D. Cadavid, O. Dobrozhan, Y. Liu, M. 2014, 105, 132103; b) D. Han, M. H. Du, C. M. Dai, D. Y. Sun, Nachtegaal, M. Zebarjadi, J. Arbiol, M. V. Kovalenko, A. S. Y. Chen, J. Mater. Chem. A 2017, 5, 6200–6210. Cabot, Nat. Commun. 2016, 7, 10766; c) S. I. Kim, K. H. Lee, [21] G. J. Snyder, A. H. Snyder, M. Wood, R. Gurunathan, B. H. H. A. Mun, H. S. Kim, S. W. Hwang, J. W. Roh, D. J. Yang, Snyder, C. Niu, Adv. Mater. 2020, 32, 2001537. W. H. Shin, X. S. Li, Y. H. Lee, G. J. Snyder, S. W. Kim, [22] C. Gayner, Y. Amouyal, Adv. Funct. Mater. 2020, 30, 1901789. Science 2015, 348, 109–114; d) K. Biswas, J. He, I. D. Blum, [23] A. Banik, B. Vishal, S. Perumal, R. Datta, K. Biswas, Energy C. I. Wu, T. P. Hogan, D. N. Seidman, V. P. Dravid, M. G. Environ. Sci. 2016, 9, 2011–2019. Kanatzidis, Nature 2012, 489, 414–418; e) Y. Liu, M. Calcabrini, [24] a) P. G. Klemens, Phys. Rev. 1960, 119, 507–509; b) C. Chang, Y. Yu, S. Lee, C. Chang, J. David, T. Ghosh, M. C. Spadaro, C. M. Ibáñez, Materials 2021, 14, 5416. Xie, O. Cojocaru-Miredin, J. Arbiol, M. Ibanez, ACS Nano [25] a) H. Wu, C. Chang, D. Feng, Y. Xiao, X. Zhang, Y. Pei, L. 2022, 16, 78–88. Zheng, D. Wu, S. Gong, Y. Chen, J. He, M. G. Kanatzidis, [4] S. Gorsse, P. Bauer Pereira, R. Decourt, E. Sellier, Chem. L. D. Zhao, Energy Environ. Sci. 2015, 8, 3298–3312; b) R. Mater. 2010, 22, 988–993. Pathak, D. Sarkar, K. Biswas, Angew. Chem. Int. Ed. 2021, 60, [5] a) C. Han, Z. Li, G. Q. Lu, S. Xue Dou, Nano Energy 2015, 15, 17686–17692; Angew. Chem. 2021, 133, 17827–17833. 193–204; b) S. Chandra, K. Biswas, J. Am. Chem. Soc. 2019, [26] J. H. Bahk, A. Shakouri, Appl. Phys. Lett. 2014, 105, 052106. 141, 6141–6145; c) S. Chandra, P. Dutta, K. Biswas, ACS Nano 2022, 16, 7–14. [6] A. Zevalkink, D. M. Smiadak, J. L. Blackburn, A. J. Ferguson, Manuscript received: May 12, 2022 M. L. Chabinyc, O. Delaire, J. Wang, K. Kovnir, J. Martin, Accepted manuscript online: July 7, 2022 L. T. Schelhas, T. D. Sparks, S. D. Kang, M. T. Dylla, G. J. Version of record online: July 21, 2022 Angew. Chem. Int. Ed. 2022, 61, e202207002 (7 of 7) © 2022 The Authors. Angewandte Chemie International Edition published by Wiley-VCH GmbH